Performance evaluation of machine-assisted interpretation of Gram stains from positive blood cultures.

Manual microscopy of Gram stains from positive blood cultures (PBCs) is crucial for diagnosing bloodstream infections but remains labor intensive, time consuming, and subjective. This study aimed to evaluate a scan and analysis system that combines fully automated digital microscopy with deep convolutional neural networks (CNNs) to assist the interpretation of Gram stains from PBCs for routine laboratory use. The CNN was trained to classify images of Gram stains based on staining and morphology into seven different classes: background/false-positive, Gram-positive cocci in clusters (GPCCL), Gram-positive cocci in pairs (GPCP), Gram-positive cocci in chains (GPCC), rod-shaped bacilli (RSB), yeasts, and polymicrobial specimens. A total of 1,555 Gram-stained slides of PBCs were scanned, pre-classified, and reviewed by medical professionals. The results of assisted Gram stain interpretation were compared to those of manual microscopy and cultural species identification by matrix-assisted laser desorption/ionization time-of-flight mass spectrometry (MALDI-TOF MS). The comparison of assisted Gram stain interpretation and manual microscopy yielded positive/negative percent agreement values of 95.8%/98.0% (GPCCL), 87.6%/99.3% (GPCP/GPCC), 97.4%/97.8% (RSB), 83.3%/99.3% (yeasts), and 87.0%/98.5% (negative/false positive). The assisted Gram stain interpretation, when compared to MALDI-TOF MS species identification, also yielded similar results. During the analytical performance study, assisted interpretation showed excellent reproducibility and repeatability. Any microorganism in PBCs should be detectable at the determined limit of detection of 105 CFU/mL. Although the CNN-based interpretation of Gram stains from PBCs is not yet ready for clinical implementation, it has potential for future integration and advancement.

Classification of fluorescent R-Band metaphase chromosomes using a convolutional neural network is precise and fast in generating karyograms of hematologic neoplastic cells.

Karyotype analysis has a great impact on the diagnosis, treatment and prognosis in hematologic neoplasms. The identification and characterization of chromosomes is a challenging process and needs experienced personal. Artificial intelligence provides novel support tools. However, their safe and reliable application in diagnostics needs to be evaluated. Here, we present a novel laboratory approach to identify chromosomes in cancer cells using a convolutional neural network (CNN). The CNN identified the correct chromosome class for 98.8% of chromosomes, which led to a time saving of 42% for the karyotyping workflow. These results demonstrate that the CNN has potential application value in chromosome classification of hematologic neoplasms. This study contributes to the development of an automatic karyotyping platform.

Two-Color Single-Molecule Tracking in Live Cells.

Measuring the kinetics of protein-protein interactions between molecules in the plasma membrane of live cells provides valuable information for understanding dynamic processes, like cellular signaling, on a molecular scale. Two-color single-molecule tracking is a fluorescence microscopy-based method to detect and quantify specific protein-protein interactions on a single-event level, providing sensitivity to heterogeneities and rare events. Fundamentally, it allows following the movement of single molecules of two different protein species in live cells with a localization precision beyond the diffraction limit of light in real time. It hence provides information about the diffusion behavior of every protein as well as about their dimerization kinetics. Here, we describe all the necessary steps to obtain two-color tracking data of plasma membrane-associated proteins in live cells using SNAP-tag and HaloTag fusion constructs and total internal reflection fluorescence (TIRF) microscopy. Also, we outline the main steps needed for analyzing the recorded data.

Transcranial magnetic stimulation for individual identification of the best electrode position for a motor imagery-based brain-computer interface.

BACKGROUND: For the translation of noninvasive motor imagery (MI)-based brain-computer interfaces (BCIs) from the lab environment to end users at their homes, their handling must be improved. As a key component, the number of electroencephalogram (EEG)-recording electrodes has to be kept at a minimum. However, due to inter-individual anatomical and physiological variations, reducing the number of electrodes bares the risk of electrode misplacement, which will directly translate into a limited BCI performance of end users. The aim of the study is to evaluate the use of focal transcranial magnetic stimulation (TMS) as an easy tool to individually optimize electrode positioning for a MI-based BCI. For this, the area of MI-induced mu-rhythm modulation was compared with the motor hand representation area in respect to their localization and to the control performance of a MI-based BCI. METHODS: Focal TMS was applied to map the motor hand areas and a 48-channel high-resolution EEG was used to localize MI-induced mu-rhythm modulations in 11 able-bodied, right-handed subjects (5 male, age: 23-31). The online BCI performances of the study participants were assessed with a single next-neighbor Laplace channel consecutively placed over the motor hand area and over the area of the strongest mu-modulation. RESULTS: For most subjects, a consistent deviation between the position of the mu-modulation center and the corresponding motor hand areas well above the localization error could be observed in mediolateral and to a lesser degree in anterior-posterior direction. On an individual level, the MI-induced mu-rhythm modulation was at average found 1.6 cm (standard deviation (SD) = 1.30 cm) lateral and 0.31 cm anterior (SD = 1.39 cm) to the motor hand area and enabled a significantly better online BCI performance than the motor hand areas. CONCLUSION: On an individual level a trend towards a consistent average spatial distance between motor hand area and mu-rhythm modulation center was found indicating that TMS may be used as a simple tool for quick individual optimization of EEG-recording electrode positions of MI-based BCIs. The study results indicate that motor hand areas of the primary motor cortex determined by TMS are not the main generators of the cortical mu-rhythm.

Applying microdroplets as sensors for label-free detection of chemical reactions.

Despite its tremendous high-throughput screening capabilities, widespread applications of droplet-based microfluidics are still limited by the poor availability of appropriate analytical assays. Here we report on a novel sensor method that exploits the osmosis-driven change in droplet size as a quantitative and label-free marker for reactions inside the droplets. We present an analysis of the underlying mechanism and apply the method for monitoring metabolic activity at a single-cell level.

[Information and participation interests of patients with depression in clinical decision making in primary care]. / Informations- und Partizipationsinteressen depressiver Patienten bei der medizinischen Entscheidungsfindung in der hausärztlichen Versorgung.

A main flaw in the primary care of depression is the patients' skepticism about adequate treatment. Despite effective treatment options patients' medical demand is poor and their information insufficient. The present study aims to improve patients' information and their participation in decision making to enhance treatment motivation and satisfaction with treatment and clinical outcome. It is unclear, though, if patients are able to show interest in information and participation because of depression-specific deficits (e.g. an impairment of cognition and concentration). In a controlled study 232 patients with depression from 30 GP practices completed questionnaires assessing information and participation needs. Patients with all severity grades of depression are highly interested in information (mean = 85.2; range 0-100) and moderately interested in participating in decision making (mean = 46.8; range 0-100). Perceived involvement in care achieved a mean of 68.4 (range 0-100). The implementation of shared decision making in the primary care of depressive patients is reasonable and will be realised and evaluated within the scope of a targeted continuing education program for general practitioners.